Introduction

The purpose of this booklet is to help those who apply and use the Tri-Blender
understand its principles of operations. With these principles the blender can be
utilized to the fullest.
Basic Operation
The Tri-Blender basically consists of a centrifugal pump head and impeller
mounted so that the normal suction port, or inlet, is pointed upward (see
illustration 1). The inlet piping consists of a patented tube-within-a-tube
arrangement. This serves to keep the liquid and dry ingredients separated
until they are in the mixing chamber. This tube-within-a-tube arrangement
eliminates one of the major problems of wet-dry mixing, that of prewetting.
The natural suction of a centrifugal pump is used to pull the powder from
the hopper through the diffuser tube.
The outer tube is used to direct the liquid into the impeller. The liquid is
introduced at the leading edge of the impeller blades where the pressure gradi-
ent of the Blender head is zero (see illustration 2). Since there is no suc-
tion at this point, liquid must be pumped to the blender by external means,
such as the supply pump (see illustration 3).
The liquid inlet is mounted tangentially. The liquid entering the blender
mixing chamber enters the chamber tangentially, in the same direction the
impeller is rotating. The impeller then strikes the liquid and accelerates it,
with a minimum of splashing. This is how the powder inlet tube can be kept
dry during operation (see illustration 4).
Should excessive splashing occur or the inlet tube plug up, the first thing
to check are impeller rotation and proper inlet tee installation. To check
for proper tee installation, trace an arrow in a continuous path from the
inlet to the outlet, without reversing direction (see illustrations 4 & 5).
In operation, the liquid being pumped into the blender is accelerated out-
ward until it strikes the blender screen. This creates a natural back pres-
sure that causes the liquid to flow downward and back toward the eye of
the impeller (see illustration 6). This forms a hollow envelope into which
the powder is introduced by the natural suction of the blender. When the
blender is operating properly the inlet tube is dry.
If the hollow envelope collapses inward problems may occur: Prewetting
which can cause lumps in the product and/or wetout which can cause plug-
ging of the inlet tube.
MAXIMUM
PRESSURE
Illustration 2
Vacuum
Illustration 1
Illustration 3
*Valves meeting 3A standards, also available
Principles of Operation
Tri-Blender

BULLETIN TBPO-97
Hopper
Valve*
Impeller
Outlet
Eye of
Impeller
Diffuser
Tube
Outer Tube
Area
Screen
Liquid
Inlet
Blending
Chamber
Illustration 5
Illustration 6
Illustration 4
Correct
Wrong
Flow and sizing of the auxiliary supply and discharge
pumps is very important for proper operation of the
blender.
Pumps are sized to maintain recommended flow at fi-
nal viscosity. In centrifugal pump systems the desired
flow, before addition of powder, is between Points A
and B, as shown on the graph. Flow should never
exceed the upper limit, Point C on the graph. Point C
is where the horsepower requirements of the blenders
will be exceeded.
If the flow would drop below Point B at final viscosity,
a noticeable drop in addition rates will occur. To cor-
rect this it will be necessary to use positive supply and
discharge pumps. This approach assures a uniform
flow rate regardless of viscosity increases.
When positive pumps are used, the flow rates are not
necessarily those which would give maximum vacuum
by the graph. By running the discharge pump faster
than the supply pump, the vacuum formed by the dis-
charge pump can be used to increase the vacuum in
the blender. The advantage of this is that smaller, less
expensive positive pumps can be used to accomplish
the same purpose.
The graph also shows the effect of temperature on the
blender vacuum. It has been found that the maxi-
mum practical limit for aqueous products is 140F
(60C). Above 140F (60C) the water will vaporize
rapidly because of the high vacuum. Vaporizing rap-
idly enough to seriously effect the vacuum as well as
Tri-Blender

Reasons why prewetting and/or

wetout may occur.
1. Inlet flow: Too much flow or
high an inlet pressure will cause
the envelope to collapse because
the blender cannot pump the liq-
uid away fast enough. It is also
possible to have too little flow
which results in a dry blender.
This is evident when powder
builds up on parts of the impel-
ler because the envelope did not
form at all.
2. High Viscosity: This causes a
natural back pressure and the
added back pressure caused by
the screen may be too much. On
higher viscosity products, the
screen is normally removed.
3. High Discharge Pressure: This may be caused by
too small a discharge line, too long a discharge line
or too high a viscosity.
The solution to these problems is proper line sizing
and the addition of a discharge pump. Even though
the Tri-Blender is basically a centrifugal pump, its
pumping ability is limited if it is going to do a good
job blending. As the back pressure increases in the
blender, the hollow envelope is forced to collapse, which
reduces its ability to blend ingredients with liquid.
Powder addition rates are difficult things to predict
because there are so many variables involved. Some of
the more important ones are:
1. Flowability
2. Moisture content
3. Humidity
4. Fat Content
5. Microscopic texture (smooth, rough)
6. Fluidity (amount of air in the product)
7. Density
8. Type of powder (granular, flake, fine,coarse, etc.)
Vacuum is one variable over which there is some
control. The amount of vacuum formed in the blender
is largely dependent on the flow and temperature.
A graph of vacuum versus flow is shown in illustration
7 for the three blenders.
Illustration 7
0
0
0
0
20 (4.5)
40 (9.1)
40 (9.1)
80 (18.2)
40 (9.1)
100 (22.7)
100 (22.7)
120 (27.3)
45 (10.2)
120 (27.3)
120 (27.3)
140 (31.8)
60 (13.6)
140 (31.8)
140 (31.8)
170 (38.6)
F2116
F3218
DS3218
F4329
10" (254mm)
20" (508mm)
24" (609.6mm)
A B
C
70F
(21C)
100F
(38C)
140F
(60C)
160F
(71C)
180F
(82C)
V
A
C

I
N
/
H
g

(
m
m
/
H
g
)
FLOW
GPM
(m
3
/hr)
wetout the internal surfaces. This wetout could
cause a buildup of powder that will plug up the inlet
tube. Temperature in excess of 160F (71C) can ac-
tually cause blow back, air will come out of the hop-
per through the powder. At 180F (82C) there is a
complete loss of vacuum.
Non-aqueous solutions may have different tempera-
ture limits depending on the vapor pressure of the liq-
uid.
Pump Sizing
The sizing of the supply and discharge pumps is very
critical to the operation of the blender and, in many
ways, more important. The addition rate of powders
and the degree of blending is very dependent on the
proper sizing of the supply and discharge pumps.
It is important to maintain a constant flow of liquid
through the blender.
A constant flow will maintain a constant vacuum.
Without a constant vacuum the powder addition rates
will not be uniform. This constant flow must be main-
tained even though the viscosity may be increasing due
to the addition of powders. Therefore, the pumps must
be sized to pump the final viscosity of the product.
The most important piece of information that needs
to be obtained is the final viscosity. If necessary,
samples can be tested at Tri-Clover to determine the
final viscosity. Other vital information required for
determining the final viscosity properly are; concen-
tration, temperature and time.
The simplest system to use is a centrifugal supply pump
and the blender, with no discharge pump. This con-
figuration is used where the final viscosity is under 500
CPS and the total discharge head is under 25 ft (7.6m).
As the viscosity approaches 500 CPS and/or the total
discharge head approaches 25 ft. (7.6m), it may be nec-
essary to add a centrifugal discharge pump.
The blender, although it is basically a centrifugal
pump, if it is used as a pump, becomes less effective as
a blender. As the back pressure increases in the blender
the liquid envelope begins to collapse. The envelope
collapses until the envelope recedes into the eye of the
impeller. This can cause prewetting and improper
mixing. Evidence of this may be in the form of fish
eye (dry centered lumps with a glossy wet surface).
In extreme cases, the normally dry powder inlet tube
may become wet. This would cause a buildup of pow-
der on the powder inlet tube I.D., even up to the but-
terfly valve. This usually results in the inlet tube be-
coming plugged.
Further evidence of insufficient discharge pumping ca-
pacity and the need for a discharge pump is shown by a
rapid addition rate on the first powder added and a gradual
slow down as the batch continues. The powder addition
may stop entirely if the extremes shown on the graph are
exceeded. In some border line cases, a centrifugal pump
on the discharge may be all that is needed to make the
system work. Usually, a duplication of the supply pump,
except for a 1" larger impeller, is all that is needed for a
discharge pump. A larger pump or a high speed pump
of slightly higher capacity may also be used.
Tri-Blender

The supply pump is always a slow speed pump

because the blender requires a flow at low pressure to
effect a good blend. This will also keep splashing to
a minimum.
Normal flow rates required at the blender for centrifu-
gal pump systems are:
F2116 40-45 GPM (9.1-10.2 m
3
/hr) @ 5-10 psi
(0.34 - 0.7 bar)
F3218 and DS3218 110-120 GPM (25.0-27.3 m
3
/hr)
@ 10-15 psi (0.7 - 1 bar)
F4329 130-140 GPM (29.5-31.8 m
3
/hr) @ 10-15
psi (0.7 - 1 bar)
These flow rates should not be exceeded or the motors
will be overloaded.
In most cases, where final viscosity is 500 CPS or higher,
positive pumps on both the supply and discharge are re-
quired. Two pumps are required to maintain constant
flow into and away from the Blender. This is necessary
because the viscosity will increase when mixing a batch.
The discharge pump is always sized to run faster than
the supply pump because; one, it has to handle a larger
volume of product, and secondly, it effects the vacuum
and, therefore, the addition rates.
It is desirable to use a vari-drive to power the discharge
pump because of this effect on vacuum. Varying the
speed allows you to adjust the vacuum for maximum
addition rates and top performance.
Normally the discharge pump is run 30 RPM faster than
the supply pump. The Tri Blender Bulletin TBM
shows the speeds normally used. These values may
vary somewhat depending on particular product char-
acteristics.
The normal flow requirements at final viscosity for the
blender in positive pump systems are:
F2116 40-45 GPM (9.1-10.2 m
3
/hr)
F3218 & DS3218 45-55 GPM (10.2-12.5 m
3
/hr)
F4329 60-70 GPM (13.6-15.9 m
3
/hr).
The Tri-Blender Bulletin TBM gives a guideline to
pump sizes. However, this may vary depending on
product characteristics and discharge heads.
Lining sizing is also important in determining heads
and flows as the viscosity increases. This also has an
important bearing on pump size.
When operating properly, the discharge pump is actu-
ally maintaining a consistant vacuum in the blender.
The discharge pump should be located close to the
blender. It should be running fast enough to pull proper
vacuum but not so fast that the pump cavitates. Evi-
dence of insufficient vacuum (and too high a product
temperature) is a puffing of air from the surface of the
powder in the hopper. Under these conditions pow-
der addition rates would be very slow or non-existent.
The use of a centrifugal supply pump and a positive
discharge pump is very hard to balance flow, particularly,
where the flow is changing as viscosity increases when
more powder is added. The system requires constant
adjustment through the use of valves making it very dif-
ficult to operate. For this reason it is NOT recommended
even though earlier Tri-Blender Bulletins specified its
use for some concentrations.
The ultimate criteria for sizing the supply and dis-
charge pumps is the final viscosity. If at all possible,
this should be obtained. Even a comparison to some-
thing else can be extremely useful. In many cases, sol-
ids content can be used to determine pump sizing,
but this can be misleading. One exception to this
would be products like stabilizers that thicken at very
low concentrations. In this case, concentrations as low
as 2% may require positive pumps. An exception to
this, on the other extreme, is sugar where viscosity is a
function of temperature and at higher temperatures
concentration in excess of 50% can be mixed with cen-
trifugal pump systems.
It all depends on the final viscosity at final tempera-
tures and if there is a time relationship to the viscosity.
Products that should be avoided are:
1. Abrasive: Products that will deteriorate seals and
impellers.
2. Effervescent: Products that emit gases will destroy
the vacuum and the powder will not flow from the
hopper.
3. High temperature: Aqueous products requiring mix-
ing over 140F because the resulting vapors will plug
up the powder tube.
4. Very high viscosity: The typical viscosity range for
Tri-Blenders is up to 25,000 cps. If viscosity ex-
ceeds 25,000 cps contact Tri-Clover.
5. Incompatible products: Products not compatible
with available seats and gaskets.
Powder Addition Rates
The three Blenders have been nominally rated as fol-
lows:
Tri-Blender

X 100 = 21.8%
100
475
Tri-Blender

F2116 50lb/min. (23kg/min.)

F3218 & DS3218 100 lb/min. (45kg/min.)
F4329 350 lb/min. (159kg/min.)
These rates are strictly estimates and do not hold true
for all powders. Such things as flowability, type of
powder, density, texture, all have a bearing on the ad-
dition rates. Also, some powders, such as stabilizers,
must be metered in slowly to get proper blending.
Addition rates of less than 50% of rated capacity can
be expected. On the other extreme, addition rates
double that of rated capacity have been obtained with
granular products such as sugar. Some products with
high fat content like chocolate powder have poor
flowability and are difficult to get out of the hopper so
addition rates are slow. Using a larger blender, with a
larger hopper opening, is required to handle these pow-
ders. Adding better flowing powders in the hopper
with these powders, when possible, often helps.
Continuous vs. Batching
The decision of using a continuous system or a batch
system is really determined by what the blender can
do. By using the flow rates required by the blender
and the powder addition rates for each blender, it can
readily be seen what the cut off point is between using
the continuous or batch method.
The F2116 requires a flow of 45 gal/min. or approxi-
mately 375 lb/min. of liquid. Using 50 lb/min. powder
addition rate the maximum, percent solids is:
The F3218 requires a flow of 120 gal/min. or approxi-
mately 1000 lb/min. of liquid. Using 100 lb/min. pow-
der addition rate the maximum percent solids is:
The F4329 requires a flow of 140 gal/min. or approxi-
mately 1164 lb/min. of liquid. Using 350 lb/min. addi-
tion rate the maximum percent solids is:
The theoretical percent solids shown above have never
been reached in actual operation. It would take ideal
X 100 = 13.3% theoretical
conditions to approach these values. There are simply
too many variables involved in the powders. In actual
practice we do not consider continuous systems for con-
centrations over 8% solids for either the F2116 or
F3218 Blenders.
Extreme care must be used in continuous systems
where precise ratios of liquid to dry are desired. If just
one lump in the powder lodged in the butterfly valve
the ratio would be upset.
Continuous systems are seldom used because of the
variations in solid content the cost of manufactured
equipment, and other difficulties of metering the pow-
der. The number of continuous systems probable to-
tal up to less than 3% of the blenders in operation
today.
If one studies the flow rates, it may appear that in posi-
tive pump systems, since the flow rates are lower, the
ratio would be higher. In a sense this is true. The
F3218 blender in a positive pump system has a flow of
55 gal./min or 457 lb./min. of liquid. If the vacuum
were essentially the same and we were to add 100 lb/
min. of powder the solids content would then be:
However, this problem occurs; at these reduced flow
rates there may not be sufficient liquid available in the
Blender head to properly absorb that much powder,
resulting in incomplete blending. In extreme cases
actual dry product may enter the positive discharge
pump and cause damage to the rotors and casings.
The solution to this is to restrict the powder addition
rates by throttling the butterfly valve on the hopper which
lowers the solids to liquid ratio. Then we are back to
essentially where we started.
The batch system is by far the best system: to assure
proper mixing, with all product incorporated, provid-
ing overall dependability.
Dual-Stage Tri-Blender - DS3218
The dual stage concept provides for double-blending,
as the product passes through the initial liquid/dry in-
gredient blending chamber directly below to a second
chamber, which effectively acts as a discharge pump.
50
375
X 100 = 10% theoretical
100
1000
X 100 = 30% theoretical
350
1164
Because of its close proximity, the dual or second cham-
ber concept actually increases, rather than decreases,
vacuum, as the vacuum from the suction of the second
pumping stage is added to that of the first stage, greatly
improving product addition rates. The double-blend
feature actually improves end product consistency, pro-
viding a smoother, more uniform blend.
With the dual stage, in applications up to 500 CPS
and up to 50 feet of discharge head, no discharge pump
is required. This is a significant improvement over
single stage units, which normally require an auxiliary
discharge pump at 25 feet of discharge head and above.
On some applications with higher viscosity a discharge
pump may be necessary.
By incorporating the discharge pump function into
the blender itself, it is possible to significantly increases
the vacuum over a wider range of process conditions.
The increased vacuum provides for fast, consistent pow-
der addition rates with minimal drop off throughout
the entire production run. This results in an overall
higher addition rate and a higher total capacity com-
pared to the single stage blender.
1997 Tri-Clover Inc. Bulletin TBPO-97 5M 7/97 Printed in the U.S.A.
Tri-Blender